The Algorithm and the Hippocratic Oath | Hacker News Reader

The Algorithm and the Hippocratic Oath

hedgehogreview.com

43 pointsblueridge18 comments

> Several years ago, I was involved in a case that illuminates the difficult position many doctors today find themselves in. The patient was pregnant, close to delivery, and experiencing dangerous declines in her baby’s heart rate. She had been on a blood thinner, which kept me, the anesthesiologist, from placing an epidural in her back. She also had strange airway anatomy, which would make it a struggle to put her to sleep quickly if an emergency cesarean section became necessary. I advised the obstetrician to perform an elective cesarean section now, in advance, while we had good working conditions, and not to wait for an emergency, where time is of the essence, and where the delay needed to induce general anesthesia might seriously injure the baby.

I am a doctor and that scenario scares me. This has a very high likelihood of stuff hitting the fan and you need to think about your plan when it does.

You want stuff to hit the fan during daytime when everyone is around. In this case, during the day surgery is around, ENT, around, other anesthesiologists all of these can rush in if needed to help you secure an airway. You also have the neonatologists around.

If it happens in the middle of the night, the staffing will be much reduced and you won’t have as many resources available.

One of the most important things to learn as a doctor is when algorithms and guidelines actually apply to the current situations.

“Life is short, the art long, opportunity fleeting, experiment treacherous, judgment difficult”

- Hippocrates

Ho bíos brakhús, hē dè tékhnē makrḗ, ho dè kairòs oxús --- O βίος βραχύς, η δὲ τέχνη μακρή, ὸ δὲ καιρὸς οξύς

Ah, it's Romanized Greek. Yes, I've forgotten most of the Classical Greek I took in college but I'm not as stupid as I sometimes feel.

https://en.m.wikipedia.org/wiki/Ars_longa,_vita_brevis

When I had a stat section late at night and the nurse who was circulating the case (for the non-medical, a “circulator” is a nurse whose job is to get whatever is needed to make the surgery happen smoothly) didn’t know how to hook up the Glidescope (um, the best airway-securing device ever) while I’m trying to mask-ventilate a full-term patient and save the baby (you don’t want to mask-ventilate highly pregnant patients; their stomachs empty slowly and they are at high risk for vomiting and then inhaling it; you want a tube straight past their vocal cords so that the lungs are protected), I went to the nurse manager on the next regular day and said that not knowing what a Glidescope is and how to set it up was an unforgivable lack of knowledge. I don’t directly blame the nurse; she was thrown into a situation she had not been trained for. I blame those who didn’t teach her before putting her on night shifts with very few other nurses around.

“This is a chance to do this case electively, in a controlled manner, in a situation in which Bad Things are monumentally more likely to occur. At noon, I can have all the help in the world. At two AM, it’s me, and I only have two hands and one brain.”

As I have said in codes before, I’m eventually out of ideas, so if you have one that we haven’t tried yet, talk. I will not judge you as dumb. I may not do it, but I will listen and consider it seriously before making that call.

Excellent article.

Star Trek proposed a magical technical Tricorder as an ultra compact futuristic little black bag. You could whip it out and aim it at someone and have scads of useful data instantly.

Reality delivered us new tech like MRIs that require their own room at a hospital. Instead of the doctor going to your home and seeing you in the context in which you live, thereby providing substantial information without having to ask, the patient now goes to the clinic.

The result: The patient has ceased to be a whole person and product of their environment in the eyes of the physician. They are merely a specimen in a petri dish.

At one time, most humans lived in small communities where everyone knew each other and the doctor was one of the smartest and most educated people there. He was in the wisdom business.

Now doctors are expected to still deliver the results of being in the wisdom business while largely being reduced to technicians who lack the wealth of contextual information that once informed their practice of medicine.

This is compounded by larger communities and mobile cultures where we no longer know everyone for years and years.

How we fix this, I don't know.

But I am reminded that House was tentatively called "Everyone Lies" in preproduction. Figuring out the lies being told in every episode was a critical part of solving the medical mystery featured.

Modern physicians no longer know x patient is an alcoholic and hiding it and y patient is a philanderer whose sickly wife stopped sleeping with him years ago and it's both rude and ineffectual to ask about such things. You will likely be lied to.

But details like that are critical to sorting out what's really wrong medically and thus how to treat it effectively. And such issues absolutely fall in the purview of medical humanities, an aspect of the profession that desperately needs rethinking to resolve what ails modern medicine and thus improve health for modern people everywhere.

Amazing comment. Ever since beginnign to experience gut issues that only appear when I am unemployed or feel my job is under threat, I have observed the ailments of friends and loved ones in a new light. You just added another layer to my questioning. To be clear, I am not reaching conclusions but just asking more questions.

Stress-induced gastritis (or even plain old stress-induced GERD) is too real and downright miserable :(

Excellent article. Practicing anesthesiologist of 25 years. Agree 100 percent with the problems highlighted in the article. These problems are increasing day by day with increasing algorithmization and protocolization of everything. So as the article mentions , there are two many individual variations and edge cases in day to day practice of medicine. An experienced doctor will have a a pretty good idea if following a protocol blindly can harm a patient but if they act on their experience and deviate from protocol they are open to liability even if the outcome was good. But I don't believe humanities education us the answer to this problem. It needs a system wide reset.

What's wrong with the simple solution, i.e., understanding when the algorithms apply and when they don't? For example, if an RCT shows that in-hospital initiation of all 4 of the heart failure GDMT's leads to reduced mortality compared to delayed initiation of GDMT, then you also have to learn that 97% of patients in the study had a baseline SBP of at least 97. Therefore, maybe you shouldn't start 4 BP-lowering meds on your HF patient with a baseline SBP of 86. If people learn when to apply rules, then I think you don't need humanities education and you don't need a system-wide reset either.

The issue isn’t the doctor knowing when to use the algorithm and when to deviate, but explaining that deviating from the algorithm was the right choice to the judge and jury in the subsequent lawsuit.

The scenario described here reminded me of https://www.astralcodexten.com/p/heuristics-that-almost-alwa.... Maybe the point is that expertise is a two way street: when you have real expert classes, you must actually trust their judgement.

The difficult edge case presented in the opening is that the standard protocol prohibits a decision that sensibly should be made. I feel the obvious fix is to amend the protocol. The time at which an operation occurs should factor into the protocol, we do prefer complex operations to occur when experts are most available. Rather than abandoning the protocol, it should be updated to reflect this. Of course there will still be cases with the protocol doesn’t handle well, but eventually those will be amended as well.

Factoring everything into the protocol would be to convert it into an exercise in discretion, which was what the protocol was designed to counteract.

I think that is what this article is addressing.

You can never have a thorough protocol that always works.

Doctors need to be trained in the limits of protocol.

This is why the humanities are important. Doctors should not just be unthinking executors of protocols. Trained human intuition, expertise, experience still matters. Knowledge of the human factors still matters.

Crucially, I don’t think the humanities are the solution here. I’d rather have a doctor that follows best practices than one who follows the supposed benefits of a humanities education — in particular I object to the idea that a doctor would follow their intuition when there is always an established best practice. The humanities are valuable, but don’t think it solves the problem statement laid out in the article.

Bioethics fall under the humanities!

A lot of it is just "being human", and you don't need separate college education for that.

Trouble is, too many doctors have internalized the ideal of being efficient robots instead.

There are lots and lots of different ways of "being human" - an infinity of ways, really, but most fall into broadly recognizeable patterns. The Humanities, properly understood, are the study of "being human" - which involves both the way you experience this (which, yeah, for most people is a learned behavior: it's hard to get outside your own perspective and evaluate your own experience), and also the way others' perspectives influence them.

There's no "just" about it. (It's like saying "Facebook is just a CRUD app, right?" - which from one point of view might be literally true, but's hardly relevant to any of the problems Meta has to solve.) Much like tech, humanities are a path of life-long learning, for which a college course of study can be (though isn't strictly necessary as) a helpful starting point, but is hardly adequate.

Efficient robots, though: yes. Many who work in tech have also internalized that mind-set.

Reply on news.ycombinator.com